Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights
Paper | Project Page | Starting with a 1D Experiment:
(optional) conda activate your_env
pip install -r requirements.txtFrom the directory containing RandOpt/:
| Step | Command |
|---|---|
| Build | docker build -f RandOpt/docker/Dockerfile_vllm -t randopt-vllm:latest . |
| Run | docker run -it --gpus all randopt-vllm:latest bash |
| Run (with data) | docker run -it --gpus all -v /path/to/RandOpt/data:/workspace/data randopt-vllm:latest bash |
Please follow the instructions CUSTOM_DATASET_GUIDE.md
First download the data here: data/README.md
Then, from the RandOpt directory:
| Mode | Command |
|---|---|
| Single node | sbatch scripts/single_node.sh |
| Multiple nodes | sbatch scripts/multiple_nodes.sh |
| Local (no Slurm) | bash scripts/local_run.sh |
Please follow the instructions baselines/README.md
@misc{gan2026neuralthickets,
title={Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights},
author={Yulu Gan and Phillip Isola},
year={2026},
eprint={2603.12228},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2603.12228},
}